Skip to main content

TextToImagePipeline

pipe(prompt, **knobs) returns the generation result: .images (one [H, W, 3] tensor each) and .seed.